There have been more than 10,000 qualified pitching seasons in history, and there’s a statistic for which the all-time leader is Marco Gonzales. I’m going to say that again: number one, out of more than 10,000.
The statistic in question? The gap between his ERA and his FIP (if you don’t know what FIP is, we’ll explain shortly). Marco currently edges out Al Maul of the 1895 Washington Senators. In 1895, Maul (who, by the way, is also in the top 10 in career stolen bases as a pitcher) sported an ERA 2.07 points better than his FIP, and Marco’s currently at a gap of 2.08. Being first would be remarkable enough, but underscoring the achievement, there’s then a 7% gap between Marco and Maul and the guy in third place. Out of more than 10,000.
While both have some predictive value, ERA and FIP are mainly descriptive stats. ERA, of course, accounts for the earned runs a pitcher has allowed, whereas FIP accounts for strikeouts, walks, HBP, and home runs and then adjusts it for that year’s offensive environment and scales it to look like ERA. In other words, it’s a measure of how well a pitcher should have performed based on the factors most within his control, which is why they call it Fielding Independent Pitching. Since runs scored are what ultimately matters, if you’re a fan of a team, you’d much rather have your pitcher outperform his FIP than the reverse, and hope to handle the anticipated regression later. Marco is doing so at a record-breaking pace.
Limiting the inquiry to the 2016 season and forward, but lowering the threshold from Qualified (one IP per game the team played) to 60 IP, there have been 969 seasons tossed by starting pitchers (as of mid-day Monday). Even with the lower threshold, Marco still comes out on top of the 969-pitcher pile, with his 2.08 overperformance 18% better than second place’s 2021 Wily Peralta.
Any time we encounter such an extreme outlier, we should first look to luck as an explanation. And indeed nobody could possibly sustain such a pace over a large sample. Of the 1,252 pitchers who’ve logged 1,000 innings in their careers, nobody’s ever beaten their career FIP by more than 0.80, and only 45 pitchers have beaten their career FIP by more than half a run.
Since FIP’s main inputs are the three true outcomes, the most obvious place to look for luck is in results on balls in play. Looking at those 969 starters with at least 60 IP since 2016—which we should to help control for era effects and to compare mid-season Marco to others who haven’t logged a full season’s worth of innings—Marco has been quite fortunate, but not historically so. His .243 BABIP allowed is .047 points luckier than the year’s league-average BABIP. That’s 54th out of the 969 pitchers in the sample, making it very lucky, but not top 5%.
Taking it a step further, though, BABIP is a rate stat, and it’ll be more impactful the more events there are. Marco is tied for the second-highest percentage of his PAs ending with a ball in play. So the advantage he’s getting from his superior BABIP relative to league average is magnified. His chance cards have given him a lot of bank errors in his favor and he keeps collecting $10 after $10. There’s been a lot of this.
The next most obvious place to look for luck in FIP-beating is in the percentage of hitters a pitcher leaves on base. Anyone following the 2022 Seattle Mariners shouldn’t need it explained to them what a big impact LOBsters can have in preventing runs. But Marco’s not getting all that lucky here. He’s stranding 5.8% more runners than the league-average starter this year, which ranks him 187th out of the 969-pitcher sample. That’s on the luckier side of the group, but it’s not even a full standard deviation away from the median.
So Marco is getting very lucky on turning balls in play into outs and a bit lucky with his strand rate. The combination of being lucky by both metrics compounds their explanatory power. But 2.08 is extreme, and I’d expect his luck to be absolutely top-tier, which it’s not. Besides, Marco’s always had some secret sauce, some version of earth’s yellow sun from which he gets superpowers: He’s beaten his Deserved Run Average (DRA), a more comprehensive, more explanatory, and more predictive metric than FIP, every year of his career, but his ERA hasn’t cratered yet.
Over a large enough sample, some amount of FIP-beating is a skill. Nobody would deny that some pitchers are better than others at controlling contact. This is why when calculating JAWS, which accounts for an entire career by Hall of Fame standards, Jay Jaffe uses bWAR, which is based on RA/9, rather than fWAR, which is based on FIP, even though Jaffe writes for FanGraphs. He’s looking at what did happen, not what perhaps should have. That’s not a dig at FIP-based WAR, by the way, which is super useful; it’s just a clear demonstration that once you reach a certain threshold, beating FIP is understood to be a skill.
And here’s the thing: Marco’s 2.08 ERA-to-FIP gap so far in 2022 is an extreme outlier, but he did the same thing in 2021. He only pitched 143.1 innings last year, which is just short of the threshold for a qualified season. But if we set the minimum threshold to 140 IP, his 1.32 difference between his FIP and ERA ranks 59th out of more than 11,000 seasons. He was luckier last year than this year, with a BABIP .052 better than league average and a LOB% 9.7% better. To get so lucky two seasons in a row is extra lucky. But it also at least suggests that there’s something Marco’s doing to earn those better results. There have been some defensive highlights, sure, but like a good 21st-Century celebrity, Marco is committed to his brand. There’s been a lot, lot, lot of this:
A common way that pitchers tend to beat FIP is by being contact-heavy pitchers who induce lots of fly balls and, specifically, tons of infield pop ups. Marco gets a decent number of infield fly balls, but he’s rarely been close to the upper echelons of the league. There’s more to it here.
One thing that defines Marco as a pitcher is his ability to mix pitches. This year, he’s throwing four pitches at least 12% of the time. And what’s more, he’ll do it in most any count, as shown by his pitch chart, where more colors means less predictability. And while no count is pure 25/25/25/25, few pitchers use this much variety. Most don’t even have four pitches at all—it’s all relative.
And this image may undersell Marco’s pitch mix. According to Aaron Goldsmith on the June 22 broadcast, Marco claims to be throwing a four-seamer way more often than Statcast is picking up. Which brings me to the next possibility to explain Marco’s success. His pitches don’t move nearly the way you’d expect them to. His sinker, for example, has about five inches less drop, and two inches less run than the average sinker, and it’s in the seventh percentile for spin rate. His curveball has four inches more drop and five inches less run than the average curveball, and it’s in the fifth percentile for spin rate. This is basically true for all of his pitches.
Some of these features are outliers for reasons we’d usually think of as good, like the extra depth on his curve, and some are what we’d usually think of as bad, like the lower spin rate. But being below average can matter less when it’s extreme. The result can be that hitters have a tough time predicting how the pitch is going to move. So Marco gets extra juice out of hitters’ uncertainty. You don’t know which pitch you’re going to get, and it won’t move the way you’re used to seeing.
If Marco was great, this would translate into whiffs and strikeouts. But Marco’s not great; he’s just making the most of what he’s got, getting hitters to make bad contact pretty often. He’s also leaning into his best pitch, his changeup, more than ever, which is a useful pitch for generating groundballs. Since groundballs score fewer runners per ball in play than other sorts of contact, that’s also helping Marco.
Of course, when it’s not enough, it’s really not enough, and hitters connect like the baseball had not kept their wives’ names out of its f’ing mouth. That’s how he’s ended up surrendering 14 home runs this year, a top-10 figure in baseball.
But since home-run rate is one of the major inputs of FIP, getting crushed when he doesn’t execute also helps explain why his FIP is so bad, expanding the gap between his ERA and FIP. As one more relevant data point, Marco’s in the top third in xwOBACON. This makes perfect sense—xwOBACON includes home runs, so he’s getting credit for a lot of weak contact, while losing credit for his worst contact.
Being first out of more than 10,000 in anything isn’t something we should expect to continue, particularly something so dependent on good fortune. A huge amount of what’s happening here is that Marco’s been getting lucky, and maybe it’s a sign that he’s bad now and his house of cards could collapse at any moment. But there’s no reason he can’t keep getting lucky, and more importantly, it’s not all luck. So maybe, just maybe, Marco can capture history and cement his status as the ultimate crafty lefty.